From Building Wheels to Embedding OpenAPI: Jingchen’s Choice of an Intelligent Ops Foundation

Background: As Jingchen’s business expanded globally, its cloud‑native systems grew in scale and complexity, leading to three structural operational challenges: (1) an incomplete global topology that separates internal service calls from external cloud resource dependencies; (2) abundant metric, log, trace, and event data that cannot be automatically correlated; and (3) high‑volume, cross‑domain alerts that require tens of minutes of manual root‑cause analysis.

Solution Overview: After technical discussions with Alibaba Cloud, Jingchen decided to stop reinventing the wheel and embed the cloud provider’s observability capabilities via OpenAPI. The solution is organized into four layers: a unified multi‑dimensional data foundation, the UModel digital‑twin for automatic topology modeling, STAROps for AI‑driven root‑cause diagnosis, and seamless OpenAPI integration into Jingchen’s self‑built SRE platform.

Unified Data Foundation (CMS 2.0)

Jingchen retained its existing RUM, Prometheus, ARMS, and SLS collectors but consolidated them under Alibaba Cloud CloudMonitor Service 2.0 (CMS 2.0). This unified platform ingests metrics, logs, traces, events, and change data into a single storage and analysis layer, eliminating the “data cannot be stitched together” problem.

UModel Digital‑Twin Topology

UModel automatically builds a full‑link topology that includes internal service calls and external cloud resources (frontend → backend → middleware → database → container). The topology updates in real time as services are deployed, scaled, or changed, providing a live “map” where each node links directly to its observability data.

STAROps Intelligent Diagnosis

Leveraging the unified data base and UModel topology, STAROps automatically traverses cross‑domain dependencies during incidents. It fetches relevant metrics, logs, and traces to produce a root‑cause analysis without requiring engineers to manually correlate timestamps. For complex cases such as frequent pod GC that span container and JVM layers, STAROps can pinpoint the exact cause.

OpenAPI Integration with the Self‑Built SRE Platform

STAROps is invoked via OpenAPI from Jingchen’s own SRE console, returning diagnostic reasoning and results as a streaming response. Engineers trigger a one‑click diagnosis and view the AI‑generated analysis within their familiar platform, preserving custom business logic while gaining cloud‑level observability.

Value Realization: By offloading topology modeling and multi‑dimensional data correlation to the cloud, Jingchen’s ops team no longer spends time maintaining heavy infrastructure. They can focus on cloud resource management and system architecture optimization, shifting from reactive fire‑fighting to proactive system health management.

Future Outlook: Jingchen plans to extend STAROps coverage to all core business systems, enabling automatic cross‑domain root‑cause linking from any entry point and further refining the OpenAPI interaction to deliver smoother, real‑time diagnostic experiences.